CN107273615A - A kind of ultra-wideband microwave humidity detection method based on machine learning - Google Patents
A kind of ultra-wideband microwave humidity detection method based on machine learning Download PDFInfo
- Publication number
- CN107273615A CN107273615A CN201710454284.5A CN201710454284A CN107273615A CN 107273615 A CN107273615 A CN 107273615A CN 201710454284 A CN201710454284 A CN 201710454284A CN 107273615 A CN107273615 A CN 107273615A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- humidity
- msubsup
- microwave
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
It is microwave detection not damaged, quick, and good portability, but face the development bottleneck of sternness in humidity range context of detection.Existing microwave MOISTURE MEASUREMENT SYSTEM carries out moisture measurement using single-frequency point mostly, and its measurement range is not high, it is difficult to use in practice.The present invention utilizes microwave attenuation principle, microwave scattering signal of the measured object of different humidity under wideband frequency is obtained by using ultra-wideband antenna, as measured object humidity regression training sample set, so as to set up measured object humidity regression model using the machine learning method for having supervision.The present invention is modeled using regression machine learning algorithm to data, and optimal training parameter is obtained using the mode of cross validation, so that the model obtained is optimized, regression error is minimum.Greatly increased the invention allows to the fabric moisture scope detected, be that the application field that goes further to of microwave detection system lays the first stone.
Description
Technical field
The present invention relates to a kind of ultra-wideband microwave humidity detection method based on machine learning, it is particularly suitable for without spoke
Penetrate, the water content of fabric is quickly and easily judged under not damaged, wide humidity range, belong to Microwave Detecting Technology field.
Background technology
At present, conventional matter water-containing rate method of testing has Oven Method, DC resistance method, capacitance method and infrared ray both at home and abroad
Method.It is non real-time online using trouble in real work although Oven Method result is accurate, and belong to and damage detection;Direct current
Resistance hygrometry is but because the D.C. resistance of measured object very big, pole plate is easy to the defects such as polarization in DC electric field and there is test
Stability is poor, error is big, the low shortcoming of versatility;And although electric capacity and infra-red method are more convenient, easily by measured object form,
The condition such as density and environment is influenceed, and belongs to contact measurement, it is difficult to which real-time online is easily measured.
Microwave wavelength scope is 1000mm to 1mm, selects suitable frequency to be penetrated into the inside of object, and different
The material bodies of interior of articles bring it about reflection of different nature, refraction, diffraction or scattering so that by handle and analyze through
The microwave signal of object, so that the characteristic for holding interior of articles is possibly realized.The advantage of Microwave Detecting Technology is:1) using non-
Ionising radiation, security of system is high, human body is not damaged, can be used with regular;2) flexible antennas and miniaturization are detected
Device causes microwave to detect that the design of tube humidity equipment has feasibility, so as to ensure that ease for use, portability is suitable for spinning
Yarn, clothes, printing and dyeing etc. industry use.
Microwave detection especially has to the detection of material internal to be widely applied very much, in recent years, the inspection of water content of materials microwave
Survey also gradually causes concern.The aqueous quantity measuring method of interior of articles based on microwave signal, it is general main using microwave transmission side
Method, mainly carries out substance moisture content measurement according to the microwave attenuation of single frequency, and single-frequency point is easily by the external world in practice
Interference, so as to cause moisture measurement error to increase.In addition, also there is detection humidity range in substance moisture content microwave detection at present
Very narrow the problem of.Current moisture Microwave Detecting Technology faces the development bottleneck of sternness in terms of investigative range in a word.
Ultra-wideband antenna is used widely in microwave detection field.UWB antennas are used as microwave sending and receiving devices, tool
Have the advantages that strong antijamming capability, simple in construction, cost are low, high, low in energy consumption with wide, message transmission rate.Ultra-wideband antenna
Transmitting and reception microwave signal, handle the ultra-wideband microwave scattered signal data received, to judge by machine learning method
The water content of measured object.The ultra wide band characteristic of antenna determines the antijamming capability of system, receives the decay effect of microwave power
Fruit has advantage than ordinary antennas.In addition, carrying out portable system the features such as ultra-wideband antenna compactedness and directionality for next step
The exploitation design of system lays the first stone.
Machine learning is a branch of artificial intelligence, and many times, almost the synonym as artificial intelligence.Letter
For list, machine learning is exactly to pass through machine learning algorithm model so that system can be from a large amount of training sample learning data point
Class or regression model, so as to do Intelligent Recognition to new sample or be given a forecast to future.Machine learning is related to can be from experience
The various algorithms automatically learnt.The basis of these algorithms is built upon mathematics and statistically, can use these
Algorithm is classified to be predicted to event, to entity, problem is diagnosed and approximation to function is modeled.
The content of the invention
It is an object of the invention to provide a kind of method that can carry out material real-time online Humidity Detection in wide humidity range.
In order to achieve the above object, the technical scheme is that there is provided a kind of ultra-wideband microwave based on machine learning
Humidity detection method, it is characterised in that comprise the following steps:
Step 1, a pair of ultra-wideband antennas are relatively fixed on the upside of testee axis, pass through vector network analyzer
Ultra-wideband impulse signal is sent to testee via a ultra-wideband antenna, is received by relative ultra-wideband antenna through tested
The ultra-wideband microwave signal of object;
Step 2, the ultra-wideband microwave signal using the testee under the method l different humidity of acquisition of step 1, by it
Humidity regression training sample is used as after normalized;
Step 3, the study generation humidity recurrence for carrying out having supervision to humidity regression training sample using machine learning algorithm
Model;
Step 4, the scattered signal to be measured using the unknown object under test of the method acquisition humidity of step 1, are returned using humidity
Return model to return scattered signal to be measured, obtain the target humidity of object under test.
Preferably, the step 2 includes:
Amplitude fading of the testee under l different humidity to specific frequency microwave is obtained, is existed in measurement testee
I-th of humidity xiUnder to the amplitude fading y of specific frequency microwaveiWhen, i=1,2 ..., l so that testee is rotated, and is surveyed respectively
Amount testee turns to microwave amplitude changing value during N number of diverse location, takes its average value wet at i-th as testee
Spend xiUnder to the amplitude fading y of specific frequency microwavei, then the point set { (x of l groups data composition is obtained1, y1) ..., (xl, yl)}。
Preferably, the step 3 includes:
Step 3.1, the point set { (x obtained according to step 21, y1) ..., (xl, yl), the standard of support vector regression problem
Formal definition is:
In formula, WTW is the factor related to model complexity;C > 0 are penalty coefficient;ε is insensitive loss region;δi、For slack variable, represent that sample deviates insensitive region ε degree;W is plane normal vector;F (w, xi) it is model output item;
B is displacement;
Step 3.2, the dual definition of support vector regression problem are:
In formula, Pij=K (xi, xj) ≡ f (w, xi);α=[α1, α2..., αl], α *=[α1 *, α2 *..., αl *];αi、
For Lagrange multiplier;
Step 3.3, the optimal solution for solving the former problem of dual problem acquisition, as regression function f (x):
In formula, K (xi, xj) it is kernel function, herein using Radial basis kernel function, then have:
K(xi, xj)=exp (- γ | | xi-x||2)
In formula, γ is the parameter of control Gaussian width;
Step 3.4, the point set { (x obtained using step 21, y1) ..., (xl, yl) by way of cross validation to punishing
Parameter γ row optimizing in penalty parameter C and kernel function, makes the mean square error MSE of recurrence minimum, passes through the optimal parameter of acquisition
Obtain regression function f (x).
Preferably, the step 4 includes:
The scattered signal to be measured of the unknown object under test of humidity is obtained using the method for step 1, is obtained using the step 3.4
To regression function f (x) scattered signal to be measured is returned, obtain the target humidity of object under test.
The present invention is based on ultra-wideband antenna, utilizes this different spy of the microwave scattering signal of the measured object through different humidity
Point, carries out the collection of ultra-wideband microwave scattered signal data to the measured object of different humidity, obtains its microwave under ultrabroad band
Through the microwave power decay of yarn reel, by means of machine learning method, feature extraction is carried out, by using such as SVMs
Learning training generation learning model is carried out to the characteristic vector of yarn reel humidity Deng machine learning, learning model is carried out using test set
Acquisition and regression effect checking, experiments verify that the accuracy of humidity regressand value is higher, meets general precision and require environment
In demand.
The present invention is originally used as the work for sending reception microwave signal in microwave detection system by the use of ultra-wideband antenna
Tool, so as to obtain in broad frequency range, decay of the different tube humidity to microwave signal under different frequency utilizes machine learning
Algorithm is trained to measured object humidity regression training, obtains after humidity regression model, measured object test sample is returned, and is carried out
The judgement of yarn reel humidity.The utilization of ultra wide band is that the moisture measurement scope of the system has obtained significant raising.The present invention is calculated
Complexity is low, and reliability is high, and system configuration flexibly, with scalability, has good in microwave detection and lossless detection field
Application prospect.
The beneficial effects of the invention are as follows:It undamaged, quick in wide humidity range, effective can be detected and sentenced with microwave
The humidity of broken yarn volume.
Brief description of the drawings
Fig. 1 is the structured flowchart of present system;
Fig. 2 a and Fig. 2 b are the structure charts using antenna;
Fig. 3 is the flow chart of SVMs processing data;
Fig. 4 CST simulation model figures;
Fig. 5 emulation data SVR training parameter selection figures;
Fig. 6 initial data and regression forecasting data comparison (emulation);
Fig. 7 regression errors (emulation);
Fig. 8 experimental data SVR training parameters selection figure;
Fig. 9 initial data and regression forecasting data comparison (experiment);
Figure 10 regression errors (experiment).
Embodiment
To become apparent the present invention, hereby with preferred embodiment, and accompanying drawing is coordinated to be described in detail below.
The present invention is based on MMU microwave measurement unit as shown in Figure 1, including is fixed on relative on the upside of testee axis
A pair of ultra-wideband antennas, send ultra-wideband impulse signal, by relative ultra wide band by vector network analyzer to testee
Antenna receives the signal through testee;The centre frequency and bandwidth of microwave signal can be set according to specific detectable substance
Put, such as present invention is directed to the detection of different humidity yarn reel, and microwave signal centre frequency is 3GHz, with a width of 4GHz.Transmitting and
Reception antenna.
A kind of ultra-wideband microwave humidity detection method based on machine learning that the present invention is provided comprises the following steps:
First, antenna yarn model is established using electromagnetic simulation software CST Microwave Studio, such as Fig. 4 institutes
Show.Using antenna structure as shown in Fig. 2 a and Fig. 2 b in emulation, w is antenna width, and L is antenna tapered length, LfFor antenna
Feed line length, h is antenna thickness, wmFor feeder line width, r1For outer oval semi-minor axis, a is outer oval major semiaxis controlling elements, rs1
For outer oval major semiaxis, b is interior oval semi-minor axis controlling elements, rs2For interior oval major semiaxis, r2For interior oval semi-minor axis.
The micro- time domain solver of solver used in emulation, microwave frequency band is 1.8GHz-5GHz.Set according to different humidity
Emulated after putting dielectric constant and electrical conductivity, altogether 40 groups of gathered data.
Then, amplitude fading of the testee under l different humidity to 1.8-5GHz microwaves is obtained, in measurement measured object
Body is in i-th of humidity xiUnder to the amplitude fading y of specific frequency microwaveiWhen, i=1,2 ..., l so that testee is rotated, point
Not Ce Liang microwave amplitude changing value of testee when turning to 16 diverse locations, take its average value as testee
I humidity xiUnder to the amplitude fading y of specific frequency microwavei, then the point set { (x of l groups data composition is obtained1, y1) ..., (xl,
yl)};
Reusing machine learning algorithm have the study of supervision to generate humidity regression model in humidity regression training sample,
Comprise the following steps:
Step 1, according to point set { (x obtained in the previous step1, y1) ..., (xl, yl), the standard of support vector regression problem
Formal definition is:
In formula, WTW is the factor related to model complexity;C > 0 are penalty coefficient;ε is insensitive loss region;δi、For slack variable, represent that sample deviates insensitive region ε degree;W is plane normal vector;F (w, xi) it is model output item;
B is displacement;
Step 2, the dual definition of support vector regression problem are:
In formula, Pij=K (xi, xj) it is kernel function;α=[α1, α2..., αl], α*=[α1 *, α2 *..., αl *];αi、For
Lagrange multiplier;
Step 3, the optimal solution for solving the former problem of dual problem acquisition, as regression function f (x):
In formula, K (xi+xj) it is kernel function, using Radial basis kernel function, then have:
K(xi, xj)=exp (- γ | | xi-x||2)
In formula, γ is the parameter of control Gaussian width;
Step 4, utilize point set { (x obtained in the previous step1, y1) ..., (xl, yl) by way of cross validation to punishment
Parameter γ in parameter C and kernel function carries out optimizing, makes the mean square error MSE of recurrence minimum, passes through the optimal parameter of acquisition
Obtain regression function f (x).
Finally, the scattered signal to be measured of the unknown object under test of humidity, the regression function obtained using the step 4 are obtained
F (x) is returned to scattered signal to be measured, obtains the target humidity of object under test.
Specific SVMs is as shown in Figure 3 to flow chart of data processing.
28 groups of data are obtained in experimentation.As shown in table 2:
The tube of table 2 humidifies experimental data
Use 80% data as training set in data handling procedure, 20% data are used as test set.Using support
Vector machine solves regression problem, it is important to the selection of kernel function and parameter, and system uses Epsilon-SVR regression models, RBF cores
Function, carries out optimizing to penalty parameter c and kernel functional parameter γ by way of cross validation, makes the mean square error MSE of recurrence
Minimum, obtains optimal parameter and sets up model, the SVR training parameters selection of emulation data and experimental data is as shown in Figure 5, Figure 8.
Then model is tested using test data.The data after data and experimental data recurrence are emulated with former data comparison as schemed
6th, shown in Fig. 9.The error that emulation data and experimental data are arrived after returning is as shown in table 3.
The test data error statistics of table 3
It can see from table, the worst error of emulation is 1.91%, and mean error is 0.85%, and standard error is
1.02%.Worst error, mean error, the standard error of experiment are respectively 2.16%, 1.75,1.78%.Emulate and test
As a result show, the use in conjunction of ultra wide band and SVMs can solve the problem that in current Humidity Detection field that detection range is narrow and ask
Topic.
Claims (4)
1. a kind of ultra-wideband microwave humidity detection method based on machine learning, it is characterised in that comprise the following steps:
Step 1, a pair of ultra-wideband antennas are relatively fixed on the upside of testee axis, by vector network analyzer via
One ultra-wideband antenna sends ultra-wideband impulse signal to testee, is received by relative ultra-wideband antenna and passes through testee
Ultra-wideband microwave signal;
Step 2, the ultra-wideband microwave signal using the testee under the method l different humidity of acquisition of step 1, by its normalizing
Humidity regression training sample is used as after change processing;
Step 3, the study generation humidity regression model for carrying out having supervision to humidity regression training sample using machine learning algorithm;
Step 4, the scattered signal to be measured using the unknown object under test of the method acquisition humidity of step 1, mould is returned using humidity
Type is returned to scattered signal to be measured, obtains the target humidity of object under test.
2. a kind of ultra-wideband microwave humidity detection method based on machine learning as claimed in claim 1, it is characterised in that institute
Stating step 2 includes:
Amplitude fading of the testee under l different humidity to specific frequency microwave is obtained, in measurement testee at i-th
Humidity xiUnder to the amplitude fading y of specific frequency microwaveiWhen, i=1,2 ..., l so that testee is rotated are measured tested respectively
Object turns to microwave amplitude changing value during N number of diverse location, takes its average value as testee in i-th of humidity xiUnder
To the amplitude fading y of specific frequency microwavei, then the point set { (x of l groups data composition is obtained1, y1) ..., (xl, yl)}。
3. a kind of ultra-wideband microwave humidity detection method based on machine learning as claimed in claim 2, it is characterised in that institute
Stating step 3 includes:
Step 3.1, the point set { (x obtained according to step 21, y1) ..., (xl, yl), the canonical form of support vector regression problem
It is defined as:
<mrow>
<mi>min</mi>
<mi>i</mi>
<mi>m</mi>
<mi>i</mi>
<mi>z</mi>
<mi>e</mi>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<msup>
<mi>W</mi>
<mi>T</mi>
</msup>
<mi>W</mi>
<mo>+</mo>
<mi>C</mi>
<munderover>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>l</mi>
</munderover>
<mrow>
<mo>(</mo>
<msub>
<mi>&delta;</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msubsup>
<mi>&delta;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>s</mi>
<mi>u</mi>
<mi>b</mi>
<mi>j</mi>
<mi>e</mi>
<mi>c</mi>
<mi>t</mi>
<mi> </mi>
<mi>t</mi>
<mi>o</mi>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>w</mi>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>b</mi>
<mo>&le;</mo>
<mi>&epsiv;</mi>
<mo>+</mo>
<msub>
<mi>&delta;</mi>
<mi>i</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>w</mi>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>b</mi>
<mo>-</mo>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>&le;</mo>
<mi>&epsiv;</mi>
<mo>+</mo>
<msubsup>
<mi>&delta;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>&delta;</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msubsup>
<mi>&delta;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>&GreaterEqual;</mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
In formula, WTW is the factor related to model complexity;C > 0 are penalty coefficient;ε is insensitive loss region;δi、For
Slack variable, represents that sample deviates insensitive region ε degree;W is plane normal vector;F (w, xi) it is model output item;B is
Displacement;
Step 3.2, the dual definition of support vector regression problem are:
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<mi>&alpha;</mi>
<mo>-</mo>
<msup>
<mi>&alpha;</mi>
<mo>*</mo>
</msup>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>&alpha;</mi>
<mo>-</mo>
<msup>
<mi>&alpha;</mi>
<mo>*</mo>
</msup>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>&epsiv;</mi>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>l</mi>
</munderover>
<mrow>
<mo>(</mo>
<msub>
<mi>&alpha;</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msubsup>
<mi>&alpha;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>)</mo>
</mrow>
<mo>+</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>l</mi>
</munderover>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&alpha;</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msubsup>
<mi>&alpha;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>s</mi>
<mi>u</mi>
<mi>b</mi>
<mi>j</mi>
<mi>e</mi>
<mi>c</mi>
<mi>t</mi>
<mi> </mi>
<mi>t</mi>
<mi>o</mi>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>l</mi>
</munderover>
<mrow>
<mo>(</mo>
<msub>
<mi>&alpha;</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msubsup>
<mi>&alpha;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mn>0</mn>
<mo>,</mo>
<mn>0</mn>
<mo>&le;</mo>
<msub>
<mi>&alpha;</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msubsup>
<mi>&alpha;</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>&le;</mo>
<mi>C</mi>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>l</mi>
</mrow>
In formula, Pij=K (xi, xj) it is kernel function;α=[α1, α2..., αl], α*=[α1 *, α2 *..., αl *];αi、For glug
Bright day multiplier;
Step 3.3, the optimal solution for solving the former problem of dual problem acquisition, as regression function f (x):
<mrow>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>l</mi>
</munderover>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msubsup>
<mi>a</mi>
<mi>i</mi>
<mo>*</mo>
</msubsup>
<mo>)</mo>
</mrow>
<mi>K</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>b</mi>
</mrow>
In formula, K (xi+xj) it is kernel function, using Radial basis kernel function, then have:
K(xi+xj)=exp (- γ | | xi-x||2)
In formula, γ is the parameter of control Gaussian width;
Step 3.4, the point set { (x obtained using step 21, y1) ..., (xl, yl) by way of cross validation to punishment parameter
Parameter γ in C and kernel function carries out optimizing, makes the mean square error MSE of recurrence minimum, is obtained by the optimal parameter of acquisition
Regression function f (x).
4. a kind of ultra-wideband microwave humidity detection method based on machine learning as claimed in claim 3, it is characterised in that institute
Stating step 4 includes:
The scattered signal to be measured of the unknown object under test of humidity is obtained using the method for step 1, is obtained using the step 3.4
Regression function f (x) is returned to scattered signal to be measured, obtains the target humidity of object under test.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710454284.5A CN107273615B (en) | 2017-06-15 | 2017-06-15 | Ultra-wideband microwave humidity detection method based on machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710454284.5A CN107273615B (en) | 2017-06-15 | 2017-06-15 | Ultra-wideband microwave humidity detection method based on machine learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107273615A true CN107273615A (en) | 2017-10-20 |
CN107273615B CN107273615B (en) | 2021-06-04 |
Family
ID=60067582
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710454284.5A Active CN107273615B (en) | 2017-06-15 | 2017-06-15 | Ultra-wideband microwave humidity detection method based on machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107273615B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112095210A (en) * | 2019-06-18 | 2020-12-18 | 株式会社岛精机制作所 | Method and system for processing driving data of knitting machine by machine learning |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102735697A (en) * | 2011-04-07 | 2012-10-17 | 中国科学院电子学研究所 | Method and apparatus for detecting deep soil humidity through microwave remote sensing |
CN203011859U (en) * | 2013-01-05 | 2013-06-19 | 开封市测控技术有限公司 | Transmission type microwave moisture detection device |
CN104036112A (en) * | 2014-04-24 | 2014-09-10 | 河海大学 | Fault diagnosis method based on support vector machine and expert system |
CN105116399A (en) * | 2015-08-27 | 2015-12-02 | 电子科技大学 | Soil humidity inversion method aiming for ultra wide band radar echo |
US20160188876A1 (en) * | 2014-12-30 | 2016-06-30 | Battelle Memorial Institute | Anomaly detection for vehicular networks for intrusion and malfunction detection |
CN106096646A (en) * | 2016-06-07 | 2016-11-09 | 衢州学院 | A kind of support vector regression model selection method |
CN106198868A (en) * | 2016-07-05 | 2016-12-07 | 深圳大学 | The method and system of Humidity Detection based on wireless aware |
CN106503288A (en) * | 2016-09-18 | 2017-03-15 | 江苏大学 | Interconnection hydraulic cylinder mechanical properties prediction algorithm based on Support Vector Machines for Regression |
CN106845544A (en) * | 2017-01-17 | 2017-06-13 | 西北农林科技大学 | A kind of stripe rust of wheat Forecasting Methodology based on population Yu SVMs |
-
2017
- 2017-06-15 CN CN201710454284.5A patent/CN107273615B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102735697A (en) * | 2011-04-07 | 2012-10-17 | 中国科学院电子学研究所 | Method and apparatus for detecting deep soil humidity through microwave remote sensing |
CN203011859U (en) * | 2013-01-05 | 2013-06-19 | 开封市测控技术有限公司 | Transmission type microwave moisture detection device |
CN104036112A (en) * | 2014-04-24 | 2014-09-10 | 河海大学 | Fault diagnosis method based on support vector machine and expert system |
US20160188876A1 (en) * | 2014-12-30 | 2016-06-30 | Battelle Memorial Institute | Anomaly detection for vehicular networks for intrusion and malfunction detection |
CN105116399A (en) * | 2015-08-27 | 2015-12-02 | 电子科技大学 | Soil humidity inversion method aiming for ultra wide band radar echo |
CN106096646A (en) * | 2016-06-07 | 2016-11-09 | 衢州学院 | A kind of support vector regression model selection method |
CN106198868A (en) * | 2016-07-05 | 2016-12-07 | 深圳大学 | The method and system of Humidity Detection based on wireless aware |
CN106503288A (en) * | 2016-09-18 | 2017-03-15 | 江苏大学 | Interconnection hydraulic cylinder mechanical properties prediction algorithm based on Support Vector Machines for Regression |
CN106845544A (en) * | 2017-01-17 | 2017-06-13 | 西北农林科技大学 | A kind of stripe rust of wheat Forecasting Methodology based on population Yu SVMs |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112095210A (en) * | 2019-06-18 | 2020-12-18 | 株式会社岛精机制作所 | Method and system for processing driving data of knitting machine by machine learning |
Also Published As
Publication number | Publication date |
---|---|
CN107273615B (en) | 2021-06-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103207377B (en) | The calibration system of partial discharge ultrahigh frequency detection device and scaling method | |
CN101487864B (en) | Production detection system used for mobile terminal antenna | |
CN107831413B (en) | The electromagnetic wave detection device and method of transmission line of electricity corona discharge based on unmanned plane | |
CN103217658A (en) | Calibration evaluating system and method of partial discharge ultrahigh-frequency detecting device based on GTEM | |
CN105227249A (en) | A kind of short wavelength emissions antenna radiation efficiency automatic monitoring method and monitoring system | |
Zhang et al. | Development of multifrequency-swept microwave sensing system for moisture measurement of sweet corn with deep neural network | |
CN107192635B (en) | The method and system of nondestructive measurement density of wood | |
CN106841838A (en) | The device and method of the cable shield status monitoring based on the measurement of frequency sweep reflection power | |
CN107273615A (en) | A kind of ultra-wideband microwave humidity detection method based on machine learning | |
CN105353236B (en) | The online remote detecting method of disk insulator state and device | |
CN204086407U (en) | A kind of field test instrument compensated based on environmental parameter | |
CN101609055B (en) | Humidity continuous measurement method based on microwave sensing technology | |
CN204964656U (en) | Portable insulator intellectual detection system appearance | |
Ala et al. | Detection of radiated EM transients by exploiting compact spherical antenna features | |
CN102508124A (en) | Detection method of partial discharge of electrical equipment and device thereof | |
Miao et al. | Location algorithm for partial discharge based on radio frequency (RF) antenna array | |
CN109596561A (en) | A kind of long-range real-time online fruit quality monitoring system and monitoring method | |
CN201194007Y (en) | Edge electric field grain moisture tester based on CAN bus | |
CN202533390U (en) | Material component content evaluator | |
CN112198472A (en) | Online remote calibration method and system for partial discharge intelligent sensor | |
CN203786054U (en) | Device for measuring humidity of papermaking felt based on microwave amplitude change | |
Lutz et al. | Microwave-based detection of internal defects in composite insulators for outdoor application | |
CN110108659A (en) | One kind being based on near-infrared reflection formula soil humidity measuring device and its detection method | |
CN204925323U (en) | Cubical switchboard partial discharge detector system | |
CN208206832U (en) | A kind of online non-destructive testing device of wheat moisture content |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |